SPATIAL STATISTICS OF TEXTONS

Gary Dahme, Eraldo Ribeiro, Mark Bush

Abstract

Texture classification is one of the most studied and challenging problems in computer vision. A key requirement of successful texture classification algorithms is their ability to quantify the complex nature and diversity of real world textures. Recent developments in automatic texture classification have demonstrated the effectiveness of representing texture elements as cluster centers of convolution responses of a filter bank. These representation of texture elements are called textons. Such methods rely primarily on similarity measurements of frequency histograms of vector quantized versions of the target texture. A main problem with these approaches is that pure frequency histograms fail to account for important spatial interaction between textons. Spatial interaction is key to classification when analyzing textures with similar texture element frequency but differ in the way the texture elements are distributed across the image. In this paper, we propose the use of co-occurrence statistics to account for the spatial interaction among texture elements. This is accomplished by calculating spatial co-occurrence statistics on the maps of textons generated by the vector quantization procedure. We demonstrate the effectiveness of our method on images from the Brodatz album as well as natural textures from a tropical pollen database. We also present a comparison with a state-of-the-art method for texture classification. Finally, our experiments show that the use of spatial statistics help improve the classification rates for certain textures that present sparse and statistically non-stationary texture elements such as pollen grain textures.

References

  1. Brodatz, P. (1966). Textures: A Photographic Album for Artists and Designers. Dover, New York.
  2. Bush, M. (2000). Deriving response matrices from central american modern pollen rain. Quaternary Research, 54:132-143.
  3. Chellapa, R., Chatterjee, S., and Bagdazian, R. (1985). Texture synthesis and compression using gaussianmarkov random fieldmodels. IEEE Trans. Systems, Man, and Cybernetics, 15:298-303.
  4. Cross, G. and Jain, A. (1983). Markov random field texture models. IEEE Trans. Pattern Analysis and Machine Intelligence, 5:25-39.
  5. Cula, O. and Dana, K. J. (2004). 3d texture recognition using bidirectional feature histograms. International Journal of Computer Vision, 59:33-60.
  6. Dong, J. and Chantler, M. J. (2005). Capture and synthesis of 3d surface texture. International Journal of Computer Vision, 62:177-194.
  7. Duda, R., P.E., H., and D.G., S. (2001). Pattern Classification. John Wiley and Sons, Inc.
  8. Haralick, R. M., Shanmugam, K., and Dinstein, I. (1973). Textural features for image classification. IEEE Trans. on Systems, Man, and Cyberneticcs, SMC-3:610-621.
  9. Leung, T. and Malik, J. (1999). Recognising surfaces using three-dimensional textons. In ICCV99, pages 1010- 1017.
  10. Leung, T. and Malik, J. (2001). Representing and recognizing the visual appearance of materials using threedimensional textons. International Journal of Computer Vision, 43:29-44.
  11. Li, S. Z. (1995). Markov random field modeling in computer vision. Springer-Verlag, London, UK.
  12. Mark, M. and Wengs, M. (2006). Introducing a new (freeware) tool for Palynology. Journal of Biogeography. (In Press).
  13. Varma, M. and Zisserman, A. (2004). A statistical approach to texture classification from single images. International Journal of Computer Vision, 62:61-81.
  14. Zhu, S. C., en Guo, C., Wang, Y., and Xu, Z. (2005). What are textons? International Journal of Computer Vision, 62(1-2):121-143. Table 4: Classification Percentage Matrix for Pollen Texture (Texton Co-occurrence). Jacobina macedoana .9 0 .1 0 0 0 0 0 0
  15. Louteridium donnelsmithii 0 .6 0 0 0 0 0 0 .4 Pachystachys lutea .1 .1 .8 0 0 0 0 0 0 Ruellia graecizans 0 0 0 .4 .3 0 .3 0 0 Tricanthera gigantea 0 0 0 .2 .8 0 0 0 0 Prenanthes alba .3 0 0 0 0 .7 0 0 0 Aleurites moluccana 0 0 0 0 0 0 1 0 0 Bocconia frutesens 0 0 0 0 0 0 0 1 0 Robinsonella mirandae 0 0 .1 0 0 0 0 0 .9 Kochia scoparia 0 0 0 0 0 0 0 0 0
  16. Table 5: Classification Percentage Matrix for Pollen Textures (Texton Frequency Histogram). Jacobina macedoana .8 0 .0 .2 0 0 0 0 0 0
  17. Louteridium donnelsmithii 0 .5 0 0 0 0 0 0 .4 .1 Pachystachys lutea 0 0 .7 .2 0 0 0 0 .1 0 Ruellia graecizans 0 0 0 .6 0 0 .4 0 0 0 Tricanthera gigantea 0 0 0 .5 .5 0 0 0 0 0 Prenanthes alba .1 0 0 .2 0 .7 0 0 0 0 Aleurites moluccana 0 0 0 0 .2 0 .8 0 0 0 Bocconia frutesens 0 0 0 0 0 0 0 .9 0 .1 Robinsonella mirandae 0 0 .4 0 0 0 0 0 .6 0 Kochia scoparia 0 .2 0 0 0 0 0 0 0 .8
Download


Paper Citation


in Harvard Style

Dahme G., Ribeiro E. and Bush M. (2006). SPATIAL STATISTICS OF TEXTONS . In Proceedings of the First International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP, ISBN 972-8865-40-6, pages 13-19. DOI: 10.5220/0001368900130019


in Bibtex Style

@conference{visapp06,
author={Gary Dahme and Eraldo Ribeiro and Mark Bush},
title={SPATIAL STATISTICS OF TEXTONS},
booktitle={Proceedings of the First International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP,},
year={2006},
pages={13-19},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0001368900130019},
isbn={972-8865-40-6},
}


in EndNote Style

TY - CONF
JO - Proceedings of the First International Conference on Computer Vision Theory and Applications - Volume 2: VISAPP,
TI - SPATIAL STATISTICS OF TEXTONS
SN - 972-8865-40-6
AU - Dahme G.
AU - Ribeiro E.
AU - Bush M.
PY - 2006
SP - 13
EP - 19
DO - 10.5220/0001368900130019